Physiological properties of rod photoreceptor electrical coupling in the tiger salamander retina
1 Cullen Eye Institute, Baylor College of Medicine, One Baylor Plaza, NC-205, Houston, TX 77030, USA
Abstract
Using dual whole-cell voltage and current clamp recording techniques, we investigated the gap junctional conductance and the coupling coefficient between neighbouring rods in live salamander retinal slices. The application of sinusoidal stimuli over a wide range of temporal frequencies allowed us to characterize the band-pass filtering properties of the rod network. We found that the electrical coupling of all neighbouring rods exhibited reciprocal and symmetrical conductivities. On average, the junctional conductance between paired rods was 500 pS and the coupling coefficient (the ratio of voltage responses of the follower cell to those of the driver cell), or K-value, was 0.07. Our experimental results also demonstrated that the rod network behaved like a band-pass filter with a peak frequency of about 2–5 Hz. However, the gap junctions between adjacent rods exhibited linearity and voltage independency within the physiological range of rods. These gap junctions did not contribute to the filtering mechanisms of the rod network. Combined with the computational modelling, our data suggest that the filtering of higher frequency rod signals by the network is largely mediated by the passive resistive and capacitive (RC) properties of rod plasma membranes. Furthermore, we found several attributes of rod electrical coupling resembling the physiological properties of gene-encoded Cx35/36 gap junctions examined in other in vitro studies. This indicates that the previously found Cx35/36 expression in the salamander rod network may be functionally involved in rod–rod electrical coupling.
, http://www.100md.com
Introduction
In the vertebrate retina, photoreceptors are electrically coupled via gap junctions (Baylor et al. 1971; Gold & Dowling, 1979; reviewed in Wu, 1994). Such coupling plays an important role in visual information processing, which improves the signal-to-noise ratio of the photoreceptor output (Lamb & Simon, 1976; Attwell et al. 1985), and increases the voltage gain of photoreceptor second-order cell synapses because these synapses have the highest gain for small signals near the photoreceptor dark potentials (Capovilla et al. 1987; Belgum & Copenhagen, 1988; Yang & Wu, 1996, 1997).
, 百拇医药
The physiological properties of rod electrical coupling have been studied extensively in the salamander retina by Attwell and colleagues (Attwell & Wilson, 1980; Attwell et al. 1984), who demonstrated that injecting a –1 nA current into a rod evoked a hyperpolarization of about 20 mV in an adjacent rod and about 4 mV in an adjacent cone. This suggests that rods are strongly electrically coupled to neighbouring rods and are weakly coupled to neighbouring cones. These data, in conjunction with computer simulations of the photoreceptor network, led to the estimation of a coupling resistance of 300 M between adjacent rods and a resistance of 5000 M between adjacent rods and cones, with the assumption that each rod was electrically coupled to four neighbouring rods and four neighbouring cones (Attwell et al. 1984). It is still not clear, however, what type(s) of gap junctions are present between rods or what the biophysical properties of rod gap junctions would be in transmitting rod signals laterally from one rod to its adjacent rods.
, http://www.100md.com
The coupled rod network has distinct temporal characteristics. In the turtle retina, Detwiler et al. (1978, 1980) first showed that the time-to-peak responses following the onset of a light flash was shorter in rods further away from a bar of light, indicating that the rod network behaves like a high-pass filter. By introducing an ‘inductive’ pathway in their model, several studies suggested that the activation of a voltage- and time-dependent inward rectifying current (Ih) (Attwell & Wilson, 1980; Hestrin, 1987) might be responsible for the rod high-pass filtering property (Detwiler et al. 1978, 1980; Torre & Owen, 1983; Attwell et al. 1984). Other studies, however, suggest that rod responses are band-pass filtered during light transmission (Attwell, 1986; Armstrong-Gold & Rieke, 2003), presumably due to the membrane capacitive property (Attwell, 1986). Nevertheless, it is uncertain whether the gap junctions between two adjacent rods would contribute to the filtering properties of the rod network, and therefore attenuate the kinetics of rod signals. Recent studies in other retinal neurones and in the neurones in other parts of the central nervous system (CNS) indicate that such a coupling pathway may behave as a low pass filter (Veruki & Hartveit, 2002a,b; Nolan et al. 1999; Galarreta & Hestrin, 2001).
, 百拇医药
Our previous study investigated the cellular localization of gap junction proteins in the salamander retina with a fish connexin35 (Cx35) antibody (that also recognizes murine connexin36, and thus named Cx35/36 in this study). Cx35 was first cloned from the fish retina (O'Brien et al. 1996), and later a homologous murine Cx36 was identified (Condorelli et al. 1998; Shl et al. 1998) and localized in the mammalian retina (Feigenspan et al. 2001; Mills et al. 2001; Deans et al. 2002). We found that Cx35/36 was localized to rod photoreceptors in the salamander retina (Zhang & Wu, 2004), where the ultrastructure of gap junctions has been identified (Custer, 1973; Mariani, 1986). Nonetheless, whether the biophysical profiles of rod electrical coupling agree with the physiological properties of Cx35/36 examined in in vitro systems (White et al. 1999; Srinivas et al. 1999; Teubner et al. 2000) is unknown. Thus it is of great interest to compare the physiology of salamander rod electrical coupling with the Cx35/36 gap junctional properties identified in in vitro (Srinivas et al. 1999; White et al. 1999; Teubner et al. 2000).
, http://www.100md.com
In order to elucidate the biophysical properties of rod electrical coupling, we used the dual whole-cell voltage- and current-clamp recording techniques to directly measure the junctional conductance and the coupling coefficient between paired rods in tiger salamander retinal slices, and also to determine at quantitative levels the strength and dynamics of rod–rod coupling. Here we show that the conductivity of amphibian rod gap junctions is smaller than previously estimated with an average conductance of about 500 pS and an average coupling coefficient (K) of 0.07. Our experimental results, combined with the modelling of the electrically coupled rod network, also suggest that rods behave like a band-pass filter and that the gap junctions between rods do not contribute to the filtering mechanisms of the rod network. In addition, we also compared several attributes of rod electrical coupling with the physiological properties of gene-encoded Cx35/36 gap junctions examined in other in vitro studies.
, http://www.100md.com
Methods
Retinal slice preparation
Larval tiger salamanders (Ambystoma tigrinum) purchased from Charles D. Sullivan, Co. (Nashville, TN, USA) and KON's Scientific Co. Inc. (Germantown, WI, USA) were used in this study. The University Committee on Animal Use at Baylor College of Medicine approved the use of animals and all animals were treated in accordance with the NIH guidelines. The dissection procedures have been previously described (Werblin, 1978; Wu, 1987). In brief, the animals were maintained on a daily 12-h light–dark cycle. The animals were decapitated and the eyes were enucleated and hemisected. The cornea, lens and vitreous were carefully removed. The retina was then removed from the posterior eyecups and was flattened, photoreceptor side up, on filter paper. The retina and filter paper were sectioned into 200–300 μm thick slices. Slices were maintained in Ringer solution at room temperature (22°C).
, http://www.100md.com
Dual whole-cell recordings
The extracellular Ringer solution consisted of (mM): 111 NaCl, 2.5 KCl, 1.8 CaCl2, 1 MgCl2, 10 dextrose, and 5 Hepes, and the pH was adjusted to 7.8 with NaOH. In some experiments, 40 mM tetraethyl ammonium chloride (TEA-Cl) replaced equimolar NaCl in the extracellular medium to block voltage-dependent outward-rectified Ik. There was no significant difference in the averaged gap junctional conductance between the two groups (with or without TEA). The electrodes were pulled from borosilicate glass (TW150F-4, World Precision Instruments, Sarasota, FL, USA) using a Flaming/Brown P-87 puller (Sutter Instrument, Novato, CA, USA), and they had a resistance of 5–7 M when filled with an intracellular solution containing the following (mM): 106 potassium gluconate, 5 NaCl, 2 MgCl2, 5 EGTA and 5 Hepes, adjusted to pH 7.4 with KOH.
, http://www.100md.com
The experiments were performed in dim room lighting. Dual whole-cell recordings were obtained from paired salamander rods with an EPC9/2 amplifier (HEKA Elektronik, Lambrecht, Germany) in voltage- or current-clamp mode. Most of the pipette capacitance was neutralized in the cell-attached configuration using the ‘Cfast’ capacitance neutralization network built into the EPC-9. After the whole-cell configuration was established, the membrane potential of both rods was initially held at –40 mV. Electrical coupling between rods was measured by applying a series of voltage or current step commands to one rod (the driver cell) in order to elicit a membrane current or voltage response in the adjacent rod (the follower cell). Current and voltage signals were filtered at 3 kHz and digitized at 2–5 kHz using a Pentium computer equipped with an ITC-16 data acquisition board (HEKA Elecktronik). Pulse software (v8.65, HEKA Elecktronik) was used to generate voltage and current commands.
, 百拇医药
Data analysis was carried out using PulseFit (v8.65, HEKA Elecktronik), Igor Pro (v4.08, WaveMetrics, Lake Oswego, OR, USA), Excel 2000 (Microsoft), and Origin (v7.0, OriginLab Corp., Northampton, MA, USA). The junctional conductance (Gj) and the coupling coefficient (K) were estimated by calculating the slope of the linear regression curve used to fit the original data. The equations were defined as follows:
and
where A was the offset of current or voltage, Ij was the transjunctional current measured from the follower cell, and Vj was the transjunctional voltage (also defined as Vj=V=Vf–Vd, Vf being the membrane voltage of the follower cell, and Vd being the membrane voltage of the driver cell). All statistic values were given as the mean ± standard deviation and n is the number of cells unless elsewhere indicated.
, http://www.100md.com
Immunocytochemistry
The salamander retinas were fixed in fresh 4% paraformaldehyde/phosphate buffered saline (PBS), pH 7.8, for 30–60 min at room temperature. Following fixation, they were rinsed extensively with PBS. For double-label experiments, the pieces of free-floating vibratome sections were blocked with 3% donkey serum in PBS with 0.5% Triton X-100–0.1% sodium azide for 2 h to overnight in order to reduce non-specific labelling. The tissues were then incubated in a mixture of primary antibodies in the presence of 1% donkey serum–PBS–0.5% Triton X-100–0.1% sodium azide for 3–5 days at 4°C. Controls lacking primary antibodies were blank. Following extensive washes with PBS containing 0.5% Triton X-100–0.1% sodium azide, immunoreactivity was revealed by overnight incubation with immunofluorescent secondary antibodies conjugated to appropriated fluorochromes. After extensive rinsing, the tissues were mounted with Vectashield (Vector Laboratories, Inc., Burlingame, CA, USA) and observed with a confocal laser-scanning microscope (Zeiss LSM 510, Carl Zeiss, Inc., Thornwood, NY, USA). Images were acquired using a 40x or 63x oil-immersion objective lens and Zeiss LSM-PC software. Intensity and size of the images were adjusted using Adobe Photoshop (v 5.0).
, http://www.100md.com
A mouse monoclonal antibody against Cx35/36 (clone 8F6.2) was obtained from Chemicon International (Temecula, CA, USA). It was used at a dilution of 1: 1000 for immunocytochemistry. A rabbit polyclonal antibody against bovine recoverin (1: 1000) was kindly provided by Dr A. M. Dizhoor (Pennsylvania College Optometry, Elkins Park, PA, USA). Secondary antibodies used in the experiments were donkey antimouse IgG and donkey antirabbit IgG conjugated to CY3 (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) or Alexa 488 (Molecular Probes, Eugene, OR, USA) and used in the dilution of 1: 100 in PBS containing 0.5% Triton X-100–0.1% sodium azide.
, 百拇医药
Results
Rod photoreceptors and their expression of Cx35/36 gap junction proteins
Our previous study has demonstrated the occurrence of Cx35/36 gap junctions in rod photoreceptors of the tiger salamander retina (Zhang & Wu, 2004). In the rod sublayer (red, labelled by recoverin antibodies) of the outer nuclear layer (ONL), Cx35/36-positive plaques (green, arrows) are present between rod cell bodies (Fig. 1A), and are restricted at the level distal to the external limiting membrane (Mariani, 1986). The plaques are colocalized with the rod fins where the ultrastructure of gap junctions was observed (Custer, 1973; Mariani, 1986; Zhang & Wu, 2004). Therefore, in order to characterize the biophysical properties of rod gap junctions, all recordings were restricted to the rod sublayer. In some cases, Lucifer yellow dye was loaded into the cells through two patch pipettes to further identify the morphology and position of the rods. One such example is shown in Fig. 1B.
, 百拇医药
A, Cx35/36 (green) immunoreactivity in the salamander outer retina. Vertical single optical section focusing at the outer nuclear layer (ONL). The retina was double labelled with recoverin antibodies (red). Rods (R, faintly stained) were located in the upper tier, whereas cones (C, strongly stained) were located in the lower tier. The characteristic round and elongated Cx35/36-positive plaques were found between rod somas in the distal portion of the ONL (arrows). Scale bar = 10 μm. B, visualization of a pair of rods simultaneously patch clamped and filled with Lucifer yellow through two recoding pipettes.
, 百拇医药
In an earlier electron microscopic study, Lasansky (1973) demonstrated that the salamander rods do not make ribbon junctional contacts to one another, suggesting that there is no excitatory chemical synaptic transmission between rods. It has also been reported that rods do not receive inhibitory inputs from horizontal cells (Lasansky, 1973; Copenhagen & Owen, 1976; Wu, 1992). Thus, our data, along with the others', suggest that rod photoreceptors are mostly interconnected by electrical synapses as opposed to chemical synapses. Therefore, most of our physiological recordings were made in the normal Ringer solution without the addition of blockers to suppress the chemical synaptic transmission.
, 百拇医药
Rod–rod junctional conductance
Simultaneous patch clamp recordings were obtained from a total of 38 pairs of rods. The mean input resistance, measured in the linear region of the I–V relationship (between –40 to –35 mV), was 275 ± 22 M (n= 71). To examine the direct current flow through the electrical synapses of paired rods, we first set the whole-cell configuration in the voltage clamp mode. The membrane potential of two adjacent rods was initially held at –40 mV, near their dark membrane potential (Attwell & Wilson, 1980). A voltage step series (V1) (from –120 to 40 mV with an increment of 20 mV) applied to one rod (driver cell) evoked current responses (I2) in the neighbouring rod (follower cell) of opposite polarity to the responses in the driver cell (I1) (Fig. 2A). This observation confirmed that the rod–rod coupling was preserved in the slice preparations. The appearance of opposite polarity of the transjunctional currents illustrated the presence of a sign-conserving electrical synapse between rod photoreceptors. The relation of transjunctional current (Ij) (measured in the follower cell) and transjunctional voltage (Vj, see Methods) at the steady state (, Fig. 2A) was approximately linear at the membrane potentials tested (Fig. 2B). Switching the driver/follower cell positions resulted in similar current responses and linear relations of Ij–Vj. This property was observed in all pairs of next-neighbouring rods.
, 百拇医药
A, the two paired rods were voltage clamped at –40 mV. When a series of voltage step commands (V1) were applied to cell1 (driver cell), the voltage-activated current responses (I1) were recorded in cell 1 (left panel), and the junctional currents of the opposite polarity (I2) were recorded in cell 2 (follower cell, right panel). B, the plot of transjunctional current (Ij) as a function of transjunctional voltage (Vj) at the steady state () seen in A. C, distribution of junctional conductance for 28 rod–rod pairs. D, the plot of the junctional conductance in each direction. Gj1,2 represents the junctional conductance measured from cell 1 coupled to cell 2, whereas Gj2,1 represents the junctional conductance measured from cell 2 coupled to cell 1. The diagonal line indicates the expected value, showing that the junctional conductance in either direction is similar.
, http://www.100md.com
To further determine the conductivity of rod coupling, we calculated the junctional conductance (Gj). This was determined by the slope of a linear regression-fitting curve (see eqn (1) in Methods) as shown in Fig. 2B. The junctional conductance of the cell in Fig. 2A was 500 pS. A histogram of junctional conductances is given in Fig. 2C. The mean junctional conductance was 500 ± 52 pS (n= 50). The comparison of coupling conductance measured in each direction (i.e. G1,2/G2,1) showed that the conductance was similar and that the electrical synaptic transmission between adjacent rods was reciprocal and symmetrical (Fig. 2D).
, http://www.100md.com
The coupling coefficient of rod–rod electrical synapses
The coupling coefficient (K) of rod–rod electrical synapses can be measured by determining the ratio of the voltage responses of the neighbouring follower cell to that of the driver cell, as shown in the rat retinal studies by Veruki & Hartveit (2002a,b). In the current clamp configuration, injecting a series of negative currents (I1, from –1000 to –100 pA with the increment of 150 pA) into a driver cell evoked a fast transient hyperpolarization followed by a characteristic peak-to-plateau depolarization (V1) in the driver cell (Fig. 3A). The voltage responses in the follower cell (V2) exhibited the same polarity as those measured in V1 (Fig. 3A). Injecting a series of positive currents (from 50 to 650 pA with the increment of 150 pA) into the same driver cell revealed a depolarization in the driver cell and a smaller depolarization in the follower cell (Fig. 3A). Note that the voltage responses of the follower cell exhibited a transient hyperpolarization, which was similar to those of the driver cell. However, the temporal property of the follower cell's transient voltage responses was slower and broader than that of the driver cell (Fig. 3B). The time-to-peak responses of the follower cell (varied with the currents injected into the driver cells) were about 20–100 ms slower than those of the driver cell. Switching the driver/follower cell positions did not change the voltage responses of the driver/follower cells. The activity observed in all pairs of next-neighbouring rods was similar. Measuring the amplitude of the voltage responses at the onset (filled symbols) or the offset (open symbols) of the current injection steps also revealed a rectified relationship between the currents injected to the driver cell and the voltage responses of the driver and follower cells at more depolarizing potentials (Fig. 3C).
, 百拇医药
A, when a series of current step commands (I1) were applied to cell 1, the voltage responses (V1) were recorded in cell 1 (left panel) and the evoked voltage responses (V2) of the same polarity were recorded in cell 2 (right panel). B, the plots of voltage responses of cell 1 and cell 2 corresponding to V1 and V2, respectively, in an expanded time scale. C, the plots of voltage responses of V1 and V2 as a function of currents injected into the driver cell (I1) at the instantaneous state (filled symbols) and the steady state (open symbols). D, the plot of the coupling coefficient in each direction of paired rods. The diagonal line indicates the expected value, showing that the coupling coefficient in either direction is similar. E, the plot of the K ratio versus the ratio of the rod input resistance (see text). The diagonal line indicates the expected value, showing that the K ratio is similar to the ratio of the input resistance.
, http://www.100md.com
The coupling coefficient (K) of rod–rod electrical transmission (see eqn (2) in Methods) measured at the instantaneous state (filled symbols, Fig. 3A) ranged from 0.02 to 0.3 with the mean value of 0.07 ± 0.01 (n= 40). Most cell pairs had symmetrical K-values for both directions (i.e. K1,2 and K2,1) (Fig. 3D). In some pairs, the coupling appeared to be rectified in spite of the fact that the junctional conductance (Gj) for each direction was symmetrical. When the mean K ratio (the larger K-value in either direction divided by the smaller K-value in the opposite direction) was calculated (2.20 ± 0.49, n= 20 cell pairs), it was found to be very close to the mean ratio of input resistances of paired rods (2.15 ± 0.45, n= 20 cell pairs) (the larger input resistance divided by the smaller input resistance). Plotting the K ratio as a function of the ratio of input resistances revealed the correlated relationship (Fig. 3E). This suggested that the rectification occurring in rod–rod coupling was due to a disparity between the membrane input resistances of the coupled rods. These results are similar to a previous study of the rat retina (Veruki & Hartveit, 2002a,b).
, 百拇医药
The voltage independency of rod–rod electrical transmission
The measurement of the gap junctional conductance demonstrated that the rod–rod coupling conductivity exhibited weak voltage sensitivity within the physiological range of rod photoreceptors (see Fig. 2B). This implies that the reciprocal communication between rods is independent of the transjunctional voltage, especially when rods are illuminated uniformly in the dark when their membrane potential is usually at –40 mV. But it is not known whether the efficient transmission between rods is affected when the resting membrane potentials of two rods differ. To test this idea, two adjacent rods were current clamped at 0 pA. As shown in Fig. 4A, a sinusoidal current pulse of 5 Hz (I1) applied to the driver cell evoked a peak-to-peak voltage response (V2) of 60 mV in the follower cell at its resting membrane potential. When the follower cell resting membrane potential was moderately hyperpolarized or depolarized by injecting a small negative or positive current, the amplitude of the follower cell voltage responses elicited by the current injection to the driver cell was less affected. However, when the resting membrane potential of the follower cell was either hyperpolarized to about –70 mV or depolarized to about –10 mV by injecting a current of ± 280 pA (I2), and the same sinusoidal current pulse of 5 Hz was applied to the driver cell, the amplitude of the follower cell voltage responses was partially suppressed. As the follower cell was hyperpolarized to more than –100 mV or depolarized to more than 0 mV, the same sinusoidal current pulse applied to the driver cell elicited much less voltage responses in the follower cell. When the normalized coupling coefficient (K) (to that determined at 0 pA holding current) was plotted as a function of the currents injected into the follower cell (Fig. 5B) and was fitted by the Gaussian curve, it showed a bell-shaped relation, reflecting that within a certain range of current injections (between –300 and 50 pA), the normalized K remained relatively constant. Outside this range, the normalized K decreased. Yet, a large amount of the residual gap junctional current still remained. The same observation was found in three out of three rod pairs.
, http://www.100md.com
A, the effect of varying membrane potentials on the rod electrical transmission. When the membrane potential of cell 2 was set at different voltages by injecting varying currents (I2), a sinusoidal current of 5 Hz (I1) applied to cell 1 elicited voltage responses in cell2 (V2). B, the plot of normalized coupling coefficient (K) (to that determined at 0 pA holding current) at each given I2. The curve was fitted with the Gaussian equation. C, the effect of varying membrane potentials on rod junctional conductance. A series of voltage step commands applied to cell 1 (V1) activated junctional currents in cell 2 (I2) at different holding potentials (V2 h: 0 mV, –40 mV and –80 mV). D, the plot of the Ij–Vj relation at each given holding potential. The conductance was determined by the slope of the Ij–Vj curve. E, the plot of the normalized conductance (to that determined at –40 mV holding potential) as the function of the holding potential of cell 2. The curve was fitted with the Boltzmann equation. The holding potential that produced half conductance was estimated at –2.66 mV.
, http://www.100md.com
A, the sinusoidal current stimuli (I1) of varying frequencies applied to cell 1 elicited voltage responses in cell 1 (V1, grey curve) and in cell 2 (V2, black curve). B, the plots of the normalized V1 (), V2 (), K () (left axis), and the degree of phase shift (, right axis) as the function of current frequency. The cut-off frequency was estimated as 33 Hz. C, the sinusoidal voltage stimuli (V1) of varying frequencies applied to cell1 activated transjunctional currents in cell 2 (I2). D, the plot of the normalized transjunctional current (I/Imax) as the function of voltage frequency.
, http://www.100md.com
Alternatively, to probe how the rod resting membrane potentials affect the rod coupling conductance, two adjacent rods were simultaneously recorded in the voltage-clamp configuration (Fig. 4C). The complete traces of current responses of the follower cell (I2) (while its membrane potential was held at 0, –40 and –80 mV) are depicted in Fig. 4C. At any given holding potential between –80 and 40 mV, the Ij–Vj relation was approximately linear (Fig. 4D). However, it should be pointed out that when the membrane potential of the follower cell was held at a potential more negative than –80 mV or more positive than 40 mV, it was difficult to measure the junctional current in the follower cells. Thus, we only determined the junctional conductance (Gj) from the slope of the Ij–Vj curve within the range of –80 to 40 mV. Then we normalized it to the conductance obtained at the –40 mV, and fitted the data points with the Boltzmann equation. We found that, at more positive holding potentials than –20 mV, the gap junctional conductance started to decrease (Fig. 4E). This finding is consistent with the results obtained in the current clamp mode. Nonetheless, while the holding potential was below –30 mV, the junctional conductance was almost independent of the rod membrane potential. The voltage that produced the half-maximal conductance was estimated to be –2.66 mV (n= 4 cell pairs). These results are consistent with above current-clamp data, suggesting that within its physiological range, the rod membrane potential would not attenuate rod electrical signals significantly.
, 百拇医药
The characteristics of frequency dependence of the rod network and gap junctions
It has been suggested that the coupled rod network acts as a band-pass filter in shaping the visual signal transmitting to the second-order cells (Attwell, 1986; Armstrong-Gold & Rieke, 2003). Yet it is not clear if the coupling pathway (or gap junctions) between rods contributes to the filtering properties of the rod network. We therefore recorded rod pairs in a whole-cell configuration that was set in the current clamp mode as shown in Fig. 5A. The sinusoidal current stimuli (I1) of varying frequencies (1–100 Hz) applied to the driver cell evoked the voltage responses of the driver cell (V1, grey curve) and the follower cell (V2, black curve). As illustrated in Fig. 5B, averaged from five rod pairs, we found that the normalized voltage response of the driver cells (V1, open triangles) was 94%, 100%, and 96% at 1 Hz, 2 Hz, and 5 Hz, respectively, whereas the normalized voltage response of the follower cell (V2, filled triangles) was 91% at 1 Hz, 100% at 2 Hz, and 87% at 5 Hz. When the frequency of the sinusoidal current stimuli (I1) increased continuously, the normalized voltage responses of both the driver and the follower cells decreased, with the latter one decreasing significantly. Similarly, the normalized coupling coefficient KN (filled circles) was 94% at 1 Hz, 100% at 2 Hz and 88% at 5 Hz, and it gradually reduced in amplitude and developed a phase-shift (open circles) over the frequency range of 5 to 100 Hz (Fig. 5B). Thus, at the low frequency (such as at 1 Hz), voltage signals of both the driver and follower cell were attenuated, and then they peaked at a frequency of about 2 Hz. In contrast, the high frequency voltage signals passing through gap junctions to the follower cell were also largely attenuated with the cut-off frequency (defined as the frequency that produced the half-maximal KN value) measured at 33 Hz. Since the attenuation of voltage responses over a wide range of temporal frequency was found not only in the follower cell but also in the driver cell, it is implied that rod voltage responses are shaped before they are transmitted to adjacent rods.
, 百拇医药
To further determine the frequency dependence of rod gap junctions, we voltage clamped the membrane potential of the follower cell (V2) at –40 mV and manipulated the membrane potential of the driver cell (V1) by applying sinusoidal voltage pulses of varying frequency. As illustrated in Fig. 5C, when the frequency of sinusoidal voltage stimuli (V1) was increased, the amplitude of the current (I2) passing through gap junctions to the follower cell did not significantly decrease over the frequency range of 1–50 Hz. Averaged from five rod pairs, we observed that, at the frequency of 50 Hz, the normalized current response was reduced by only 10 ± 3% of the maximum current measured at the frequency of 1 Hz (Fig. 5D). In contrast, the normalized KN at the frequency of 50 Hz was reduced by 57 ± 10% of the maximum KN determined at the frequency of 2 Hz. Therefore, the measurement of current responses as a function of the frequency was in direct contradiction to the measurement of voltage responses as a function of the frequency. This suggests that the higher frequency filtering property observed here most likely reflects the RC property of the rod network.
, 百拇医药
In order to confirm whether rod gap junctions behave as a resistor, and to examine how RC properties of the rod membrane are related to the filtering properties, we analysed a simplified model of the rod network (Fig. 6A). In this model, the membrane resistance (R1 (or 1/G1) and R2 (or 1/G2)) and capacitances (C) represent the lumped equivalent of the resistance and capacitance across driver (1) and follower (2) rod membranes, while two rods are connected through a resistive pathway Rj (= 1/Gj). Assuming Ih is ignored, and therefore the high pass filtering pathway is not considered here, the equation defining the relationship of coupling coefficient (K) as the function of frequency can be expressed as
, http://www.100md.com
At a given capacitance of 40 pF (Attwell et al. 1984) and an input resistance of 250 M (obtained from this study), the amplitude of rod signals spreading to the adjacent rods is frequency dependent (Fig. 6B). The cut-off frequency calculated from this model is 32 Hz (*, Fig. 6C), which is very close to what we measured experimentally. In addition, considering the rod input resistance variation from 100 to 500 M, the normalized voltage responses at a given higher frequency (to the steady state responses) would be attenuated significantly. The gap junctional conductance affecting the frequency-dependent voltage attenuation would be much less spectacular than that of rod membrane input resistance (Fig. 6B and C). Therefore, the agreement of the physiological data and the mathematical calculations supports the notion that rod gap junctions behave in a linear (ohmic) manner in mediating rod–rod electrical coupling. The frequency-dependent attenuation of rod voltage responses is not surprising given the rod plasma membrane capacitive and resistive properties.
, 百拇医药
A, a schematic diagram of the equivalent circuit of the rod pair. The two next neighbouring rods are interconnected by a resistance of Rj (= 1/Gj). R1, R2 and C represent the lumped equivalent of the resistance and capacitance across each rod membrane in the pairs, where R1= 1/G1 and R2= 1/G2 (G1 and G2 are the conductance of cell 1 and cell 2, respectively). B, the computer simulated normalized coupling coefficient (KN) changes as a function of frequency (when Gj and R2 were manipulated using different values). C, the plot of cut-off frequency (see eqn (3) in Results) as the function of junctional conductance at each given membrane input resistance. The calculated cut-off frequency is 32 Hz (*) when Gj, R2, and C are 500 pS, 250 M, and 40 pF, respectively.
, http://www.100md.com
The intensity dependence of the coupled rod network
In the salamander retina, rods are able to respond to light ranging in intensity from a few photons to several thousands of photons with characteristic temporal kinetics (Bader et al. 1978; Attwell et al. 1984; Yang & Wu, 1997). While rods respond to dim light with a slow graded hyperpolarization, their responses to brighter light become faster with a typical transient hyperpolarization. It is likely that when rod signals propagate through the rod network, the amplitude of rod voltage responses would not only be limited by the coupling strength of the electrical synapses, but would also be attenuated by the filtering mechanism as the light intensity increases. In order to determine the degree of filtering-induced attenuation at different log units of intensity, we measured the time-to-peak of rod transient responses obtained from intracellular recordings (n= 3) (see Fig. 3, Yang & Wu, 1997; and X. L. Yang & S. M. Wu, unpublished results), and calculated the amplitude of transient voltage responses of the follower cell with/without considering RC properties of the rod membrane. We found that during the brightest light flash, the rise time of this transient response could be less than 50 ms. Our calculations showed that over eight log units of the light intensity span, the dim light responses (log I < (4)) of rods with slow kinetics would efficiently pass through rod gap junctions (, Fig. 7). However, as light intensity increases, the transient rod responses would be brief enough to be filtered (Fig. 5). Therefore, the peak amplitude of rod responses (log I > (4)) would likely be attenuated (, frequency, Fig. 7). Thus the light responses of rods laterally spreading within the network would be intensity dependent.
, http://www.100md.com
The plots of the voltage-intensity relation of the driver cell (), the follower cell without considering filtering properties of the rod network (), and the follower cell having peak responses filtered by the rod network () as light intensity increases.
Discussion
The anticipation of Cx35/36 gap junctions in rod electrical coupling
Since we have previously demonstrated that Cx35/36 gap junction proteins are present in an organized pattern in the rod network of the salamander retina (Zhang & Wu, 2004), one aim of this continuing study is to understand whether the physiological property of rod electrical coupling is consistent with the profiles of Cx35/36 gap junctions. Although, to date, the lack of potent and specific Cx35/36 gap junction blockers makes it difficult to pharmacologically distinguish Cx35/36 from other connexin proteins in salamander rods, there are several lines of evidence suggesting that Cx35/36 may play an important role in the rod electrical coupling of this species. First, morphologically, we have shown that among all of the connexin proteins examined, only Cx35/36 is exclusively localized in rod photoreceptors at locations where ultrastructure of gap junctions has been identified (Custer, 1973; Mariani, 1986; Zhang & Wu, 2004). Secondly, the staining pattern of Cx35/36 in the rod network is similar to that of Cx35/36 gene-encoded channels expressed in transfected human HeLa cells (Teubner et al. 2000).
, http://www.100md.com
Thirdly, in the present electrophysiological study, we have demonstrated that the biophysical properties of rod gap junctions resemble the physiology of gene-encoded Cx35/36 gap junctions. Several earlier reports have shown that Cx35/36 gap junctions studied in in vitro expression systems exhibit three characteristics: (1) small channel conductance; (2) voltage independence; and (3) formation of homologous channels (Srinivas et al. 1999; White et al. 1999; Teubner et al. 2000). Similarly, in the salamander retina, the rod coupling exhibited the smaller conductance of 500 pS. Within the rod physiological range, the conductivity of rod gap junctions was independent of the transjunctional voltage between rods, and was also independent of the rod plasma membrane potential. Furthermore, since current flow passed equally well in both directions, this suggests that rod electrical synapses were symmetrical and bi-directional. These are results consistent with the notion that Cx35/36 forms homologous gap junctions between neighbouring neurones (White et al. 1999; Teubner et al. 2000). Interestingly, all of these characteristics have also been observed in Cx35/36-containing neurones studied in in situ brain slices (Landisman et al. 2002; Galarreta & Hestrin, 1999) and in Cx35/36-positive AII amacrine/ON cone bipolar cells studied in rat retinal slices (Veruki & Hartveit, 2002a,b). Therefore, our data indicate that Cx35/36 gap junctions are likely to mediate rod–rod coupling in the salamander retina. However, as selective gap junctional blockers are discovered, further pharmacological and physiological studies are needed to confirm this claim.
, http://www.100md.com
Small gap junctional conductivity mediating salamander rod electrical coupling
Our characterization of rod gap junctional properties revealed a weak coupling of rod photoreceptors. The experimentally measured junctional conductance of 500 pS was much smaller than the previously computer-simulated value of 3.3 nS (Attwell et al. 1984). Yet, it is much closer to the cone–cone coupling conductance found in the mammalian retina (Hornstein et al. 2004; Li & DeVries, 2004). It may be argued that in the slice preparations, the rod network was partially cut off, especially for superficial cell pairs, and that the junctional conductance in vivo might be larger than what we observed here. However, Attwell & Wilson (1980) found that rods were coupled more strongly in the retinal slices than in the flat-mount retina. They speculated that this might be due to the decrease of the number of paths available for current flow away from the site of current injection. Thus the slice preparation itself should not be a significant factor affecting the small rod junctional conductance.
, http://www.100md.com
We have quantitatively studied gap junction patterns between rods. We previously calculated the average number of Cx35/36 positive plaques between paired rods to be 4.63 (Zhang & Wu, 2004). Assuming the unitary channel conductance of Cx35/36 is 10–14 pS (Srinivas et al. 1999) and the coupling conductance between paired rods is 500 pS, we estimated that there should be as many as 35–50 channels (500 x 10–12)/(10–14 x 10–12)) that are open under dark-adapted conditions. Each Cx35/36-positive plaque may therefore accommodate 8–11 activated gap junctions. In a freeze fracture study of gap junctions in the toad retina, Gold & Dowling (1979) showed that the average area per junction in rod–rod pairs was 0.15 ± 0.05 μm2 and the density of junctional particles was 5 x 103μm–2. This suggests that the number of junctional particles in a plaque would be 750 (0.15 x 5 x 103). This also suggests that the activation of 8–11 gap junctions in the dark only represents 1–1.5% (8/750 or 11/750) of the total number of channels in a cluster. This percentage is much smaller than the estimate of 10% of the total number of channels in a cluster assumed to be activated as suggested in a Cx43 study by Bukauskas et al. (2000).
, 百拇医药
The low coupling conductance (500 pS) resulted in a small coupling coefficient of 0.07, which indicates a weak coupling strength between rods. This weak interconnection between rods may be desirable, because the greatest sensitivity of rods is at the effective operational range, i.e. near the dark membrane potential (Attwell & Wilson, 1980; Capovilla et al. 1987; Yang & Wu, 1996, 1997), and the weak coupling ensures a given signal divided into smaller signals that will fall within the sensitive and linear operational range of rods. These small signals are then expected to be transmitted/amplified precisely and maximally to the second-order cells by means of elaborate chemical synapses. Likewise, in the CNS (including neocortex, cerebellum and thalamus), this similar coupling coefficient has been shown to depolarize neighbouring interneurones to a subthreshold potential, thereby facilitating the synchronized firing of action potentials (reviewed in Galarreta & Hestrin, 2001). Therefore, we think that the small conductivity of rod gap junctions may be necessary to maintain rods in the optimal state for integrating incoming small signals within a narrow operational range.
, 百拇医药
The significance of the linearity of rod gap junctions
The voltage independence of rod gap junctions within the rod physiological range combined with the frequency independence of rod junctional currents strongly implies that the rod–rod coupling pathway behaves as a linear resistor. This feature may provide a mechanism to prevent rod uncoupling within the rods' dynamic range and also to improve the signal-to-noise ratio within the whole intensity range of light illumination. While voltage would not be the primary modulator of rod gap junctions, it is possible that they would be modulated by neuromodulators and/or intracellular second messengers. In other electrically coupled retinal neurones, such as horizontal and amacrine cells, dopamine, cAMP, and nitric oxide have been found to modulate the gap junctional conductivity of these cells (Lasater, 1987; Hampson et al. 1992; Mills & Massey, 1995; Xin & Bloomfield, 2000). Correspondingly, this may be true in rod electrical coupling. A recent report showed that Cx35 hemi-channels expressed in Xenopus oocytes were modulated by the cAMP-mediated protein kinase A (PKA) pathway (Mitropoulou & Bruzzone, 2003). The consensus sequence sites for PKA phosphorylation have been identified in Cx35 proteins (O'Brien et al. 1996). Furthermore, the dye diffusion coefficient in cells expressing Cx35 was modulated by PKA (O'Brien et al. 2004), and Cx36 associated proteins were found to be phosphorylated by cAMP (Sitaramayya et al. 2003). Therefore, if Cx35/36 is functionally involved in rod electrical coupling, rod electrical synapses are anticipated to be modulated at the cellular and molecular levels. These synapses may also be targeted by second messenger mediated-signalling transduction pathways or by light-dependent signalling pathways. Further pharmacological and physiological studies of the modulation of rod–rod coupling are needed in order to confirm this hypothesis.
, http://www.100md.com
The temporal filtering property of the rod network shapes intensity-dependent rod signals
In this study, as shown in Fig. 6, we demonstrated that the low-pass filtering of higher frequency rod signals might be attributed to the RC properties of the rod plasma membrane. This conclusion added a piece of additional evidence to the mechanisms of the band-pass filtering property of the rod network (Detwiler et al. 1978, 1980; Torre & Owen, 1983; Attwell et al. 1984). We found that voltage responses of coupled rods were gradually attenuated as the frequency of current stimuli increased beyond 5 Hz. This indicates that rod electrical synapses favour the transmission of slow potential changes and it agrees with the previous findings of frequency dependence of Cx35/36-containing retinal neurones and brain interneurones (Veruki & Hartveit, 2002a,b; Galarreta & Hestrin, 1999; Nolan et al. 1999; Landisman et al. 2002). One possible explanation for this finding is that Cx35/36 gap junctions might impart a ‘low-pass’ filtering property (Veruki & Hartveit, 2002a,b; Nolan et al. 1999; Galarreta & Hestrin, 2001). However, we found that the transjunctional current under voltage clamp conditions was not significantly altered as the frequency increased (Fig. 5C), indicating that DC coupling is attributed to rod gap junctions. Thus, our data argue that the attenuation of voltage responses of coupled rods by signal frequencies is not the result of the frequency-dependent processes in the gap junctions between rods. Instead it reflects rod membrane RC filtering properties. This argument is supported by our mathematical calculations in which we showed that the cut-off frequency (32 Hz) was very close to the value (33 Hz) measured experimentally, which is a result consistent with the notion that gap junctions in salamander rods behave in a linear manner. A similar observation was also made in the study of primate cone–cone coupling (Hornstein et al. 2004).
, http://www.100md.com
Rod photoreceptors are non-spiking neurones in that they respond to light with graded hyperpolarization (Schwartz, 1976; Yang & Wu, 1997). The current response to a single photon is about 2 pA with a rise time of about 2 s (Baylor et al. 1979; Baylor & Nunn, 1986). When light becomes brighter, rod light responses become larger with a faster rise time (Bader et al. 1978; Yang & Wu, 1997). For a saturating light flash, the peak current response is about 50–100 pA and with a rapid rise time of 20–50 ms (Baylor & Nunn, 1986). The frequency-dependent attenuation of the coupled rod network (Fig. 7) allows small signals with slow kinetics to spread to the neighbouring rod without much attenuation, whereas large responses with faster kinetics could be attenuated. This may be a mechanism used by the retina to suppress rod inputs to second-order cells during bright light so that the cone input can take a more dominant role.
, 百拇医药
References
Armstrong-Gold CE & Rieke F (2003). Bandpass filtering at the rod to second-order cell synapse in salamander (Ambystoma tigrinum) retina. J Neurosci 23, 3796–3806.
Attwell D (1986). The Sharpey-Schafer lecture. Ion channels and signal processing in the outer retina. Q J Exp Physiol 71, 497–536.
Attwell D & Wilson M (1980). Behaviour of the rod network in the tiger salamander retina mediated by membrane properties of individual rods. J Physiol 309, 287–315.
, 百拇医药
Attwell D, Wilson M & Wu SM (1984). A quantitative analysis of interactions between photoreceptors in the salamander (Ambystoma) retina. J Physiol 352, 703–737.
Attwell D, Wilson M & Wu SM (1985). The effect of light on the spread of signals through the rod network of the salamander retina. Brain Res 343, 79–88.
Bader CR, MacLeish PR & Schwartz EA (1978). Responses to light of solitary rod photoreceptors isolated from tiger salamander retina. Proc Natl Acad Sci U S A 75, 3507–3511.
, http://www.100md.com
Baylor DA, Fuortes MG & O'Bryan PM (1971). Receptive fields of cones in the retina of the turtle. J Physiol 214, 265–294.
Baylor DA, Lamb TD & Yau KW (1979). Responses of retinal rods to single photons. J Physiol 288, 613–634.
Baylor DA & Nunn BJ (1986). Electrical properties of the light-sensitive conductance of rods of the salamander Ambystoma tigrinum. J Physiol 371, 115–145.
Belgum JH & Copenhagen DR (1988). Synaptic transfer of rod signals to horizontal and bipolar cells in the retina of the toad (Bufo marinus). J Physiol 396, 225–245.
, http://www.100md.com
Bukauskas FF, Jordan K, Bukauskiene A, Bennett MV, Lampe PD, Laird DW & Verselis VK (2000). Clustering of connexin 43-enhanced green fluorescent protein gap junction channels and functional coupling in living cells. Proc Natl Acad Sci U S A 97, 2556–2561.
Capovilla M, Hare WA & Owen WG (1987). Voltage gain of signal transfer from retinal rods to bipolar cells in the tiger salamander. J Physiol 391, 125–140.
Condorelli DF, Parenti R, Spinella F, Trovato Salinaro A, Belluardo N, Cardile V & Cicirata F (1998). Cloning of a new gap junction gene (Cx36) highly expressed in mammalian brain neurons. Eur J Neurosci 10, 1202–1208.
, 百拇医药
Copenhagen DR & Owen WG (1976). Coupling between rod photoreceptors in a vertebrate retina. Nature 260, 57–59.
Custer NV (1973). Structurally specialized contacts between the photoreceptors of the retina of the axolotl. J Comp Neurol 151, 35–56.
Deans MR, Volgyi B, Goodenough DA, Bloomfield SA & Paul DL (2002). Connexin36 is essential for transmission of rod-mediated visual signals in the mammalian retina. Neuron 36, 703–712.
, 百拇医药
Detwiler PB, Hodgkin AL & McNaughton PA (1978). A surprising property of electrical spread in the network of rods in the turtle's retina. Nature 274, 562–565.
Detwiler PB, Hodgkin AL & McNaughton PA (1980). Temporal and spatial characteristics of the voltage response of rods in the retina of the snapping turtle. J Physiol 300, 213–250.
Feigenspan A, Teubner B, Willecke K & Weiler R (2001). Expression of neuronal connexin36 in AII amacrine cells of the mammalian retina. J Neurosci 21, 230–239.
, 百拇医药
Galarreta M & Hestrin S (1999). A network of fast-spiking cells in the neocortex connected by electrical synapses. Nature 402, 72–75.
Galarreta M & Hestrin S (2001). Electrical synapses between GABA-releasing interneurons. Nat Rev Neurosc 2, 425–433.
Gold GH & Dowling JE (1979). Photoreceptor coupling in retina of the toad, Bufo marinus. I. Anatomy. J Neurophysiol 42, 292–310.
Hampson EC, Vaney DI & Weiler R (1992). Dopaminergic modulation of gap junction permeability between amacrine cells in mammalian retina. J Neurosci 12, 4911–4922.
, http://www.100md.com
Hestrin S (1987). The properties and function of inward rectification in rod photoreceptors of the tiger salamander. J Physiol 390, 319–333.
Hornstein EP, Verweij J & Schnapf JL (2004). Electrical coupling between red and green cones in primate retina. Nat Neurosci 7, 745–750.
Lamb TD & Simon EJ (1976). The relation between intercellular coupling and electrical noise in turtle photoreceptors. J Physiol 263, 257–286.
, http://www.100md.com
Landisman CE, Long MA, Beierlein M, Deans MR, Paul DL & Connors BW (2002). Electrical synapses in the thalamic reticular nucleus. J Neurosci 22, 1002–1009.
Lasansky A (1973). Organization of the outer synaptic layer in the retina of the larval tiger salamander. Philos Trans R Soc Lond B Biol Sci 265, 471–489.
Lasater EM (1987). Retinal horizontal cell gap junctional conductance is modulated by dopamine through a cyclic AMP-dependent protein kinase. Proc Natl Acad Sci U S A 84, 7319–7323.
, 百拇医药
Li W & DeVries SH (2004). Separate blue and green cone networks in the mammalian retina. Nat Neurosci 7, 75175–75176.
Mariani AP (1986). Photoreceptors of the larval tiger salamander retina. Proc R Soc Lond B Biol Sci 227, 483–492.
Mills SL & Massey SC (1995). Differential properties of two gap junctional pathways made by AII amacrine cells. Nature 377, 734–737.
Mills SL, O'Brien JJ, Li W, O'Brien J & Massey SC (2001). Rod pathways in the mammalian retina use connexin 36. J Comp Neurol 436, 336–350.
, 百拇医药
Mitropoulou G & Bruzzone R (2003). Modulation of perch connexin35 hemi-channels by cyclic AMP requires a protein kinase A phosphorylation site. J Neurosci Res 72, 147–157.
Nolan MF, Logan SD & Spanswick D (1999). Electrophysiological properties of electrical synapses between rat sympathetic preganglionic neurones in vitro. J Physiol 519, 753–764.
O'Brien J, Al-Ubaidi MR & Ripps H (1996). Connexin 35: a gap-junctional protein expressed preferentially in the skate retina. Mol Biol Cell 7, 233–243.
, http://www.100md.com
O'Brien J, Nguyen HB & Mills SL (2004). Cone photoreceptors in bass retina use two connexins to mediate electrical coupling. J Neurosci 24, 5632–5642.
Schwartz EA (1976). Electrical properties of the rod syncytium in the retina of the turtle. J Physiol 257, 379–406.
Sitaramayya A, Crabb JW, Matesic DF, Margulis A, Singh V, Pulukuri S & Dang L (2003). Connexin 36 in bovine retina: lack of phosphorylation but evidence for association with phosphorylated proteins. Vis Neurosci 20, 385–395.
, 百拇医药
Shl G, Degen J, Teubner B & Willecke K (1998). The murine gap junction gene connexin36 is highly expressed in mouse retina and regulated during brain development. FEBS Lett 428, 27–31.
Srinivas M, Rozental R, Kojima T, Dermietzel R, Mehler M, Condorelli DF, Kessler JA & Spray DC (1999). Functional properties of channels formed by the neuronal gap junction protein connexin36. J Neurosci 19, 9848–9855.
Teubner B, Degen J, Sohl G, Guldenagel M, Bukauskas FF, Trexler EB, Verselis VK, De Zeeuw CI, Lee CG, Kozak CA, Petrasch-Parwez E, Dermietzel R & Willecke K (2000). Functional expression of the murine connexin 36 gene coding for a neuron-specific gap junctional protein. J Membr Biol 176, 249–262.
, 百拇医药
Torre V & Owen WG (1983). High-pass filtering of small signals by the rod network in the retina of the toad, Bufo marinus. Biophys J 41, 305–324.
Veruki ML & Hartveit E (2002a). AII (Rod) amacrine cells form a network of electrically coupled interneurons in the mammalian retina. Neuron 33, 935–946.
Veruki ML & Hartveit E (2002b). Electrical synapses mediate signal transmission in the rod pathway of the mammalian retina. J Neurosci 22, 10558–10566.
, 百拇医药
Werblin FS (1978). Transmission along and between rods in the tiger salamander retina. J Physiol 280, 449–470.
White TW, Deans MR, O'Brien J, Al-Ubaidi MR, Goodenough DA, Ripps H & Bruzzone R (1999). Functional characteristics of skate connexin35, a member of the gamma subfamily of connexins expressed in the vertebrate retina. Eur J Neurosci 11, 1883–1890.
Wu SM (1987). Synaptic connections between neurons in living slices of the larval tiger salamander retina. J Neurosci Meth 20, 139–149.
, http://www.100md.com
Wu SM (1992). Feedback connections and operation of the outer plexiform layer of the retina. Curr Opin Neurobiol 2, 462–468.
Wu SM (1994). Synaptic transmission in the outer retina. Annu Rev Physiol 56, 141–168.
Xin D & Bloomfield SA (2000). Effects of nitric oxide on horizontal cells in the rabbit retina. Vis Neurosci 17, 799–811.
Yang XL & Wu SM (1996). Response sensitivity and voltage gain of the rod- and cone-horizontal cell synapses in dark- and light-adapted tiger salamander retina. J Neurophysiol 76, 3863–3874.
, http://www.100md.com
Yang XL & Wu SM (1997). Response sensitivity and voltage gain of the rod- and cone-bipolar cell synapses in dark-adapted tiger salamander retina. J Neurophysiol 78, 2662–2673.
Zhang J & Wu SM (2004). Connexin35/36 gap junction proteins are expressed in photoreceptors of the tiger salamander retina. J Comp Neurol 470, 1–12., 百拇医药(Jian Zhang and Samuel M. )
Abstract
Using dual whole-cell voltage and current clamp recording techniques, we investigated the gap junctional conductance and the coupling coefficient between neighbouring rods in live salamander retinal slices. The application of sinusoidal stimuli over a wide range of temporal frequencies allowed us to characterize the band-pass filtering properties of the rod network. We found that the electrical coupling of all neighbouring rods exhibited reciprocal and symmetrical conductivities. On average, the junctional conductance between paired rods was 500 pS and the coupling coefficient (the ratio of voltage responses of the follower cell to those of the driver cell), or K-value, was 0.07. Our experimental results also demonstrated that the rod network behaved like a band-pass filter with a peak frequency of about 2–5 Hz. However, the gap junctions between adjacent rods exhibited linearity and voltage independency within the physiological range of rods. These gap junctions did not contribute to the filtering mechanisms of the rod network. Combined with the computational modelling, our data suggest that the filtering of higher frequency rod signals by the network is largely mediated by the passive resistive and capacitive (RC) properties of rod plasma membranes. Furthermore, we found several attributes of rod electrical coupling resembling the physiological properties of gene-encoded Cx35/36 gap junctions examined in other in vitro studies. This indicates that the previously found Cx35/36 expression in the salamander rod network may be functionally involved in rod–rod electrical coupling.
, http://www.100md.com
Introduction
In the vertebrate retina, photoreceptors are electrically coupled via gap junctions (Baylor et al. 1971; Gold & Dowling, 1979; reviewed in Wu, 1994). Such coupling plays an important role in visual information processing, which improves the signal-to-noise ratio of the photoreceptor output (Lamb & Simon, 1976; Attwell et al. 1985), and increases the voltage gain of photoreceptor second-order cell synapses because these synapses have the highest gain for small signals near the photoreceptor dark potentials (Capovilla et al. 1987; Belgum & Copenhagen, 1988; Yang & Wu, 1996, 1997).
, 百拇医药
The physiological properties of rod electrical coupling have been studied extensively in the salamander retina by Attwell and colleagues (Attwell & Wilson, 1980; Attwell et al. 1984), who demonstrated that injecting a –1 nA current into a rod evoked a hyperpolarization of about 20 mV in an adjacent rod and about 4 mV in an adjacent cone. This suggests that rods are strongly electrically coupled to neighbouring rods and are weakly coupled to neighbouring cones. These data, in conjunction with computer simulations of the photoreceptor network, led to the estimation of a coupling resistance of 300 M between adjacent rods and a resistance of 5000 M between adjacent rods and cones, with the assumption that each rod was electrically coupled to four neighbouring rods and four neighbouring cones (Attwell et al. 1984). It is still not clear, however, what type(s) of gap junctions are present between rods or what the biophysical properties of rod gap junctions would be in transmitting rod signals laterally from one rod to its adjacent rods.
, http://www.100md.com
The coupled rod network has distinct temporal characteristics. In the turtle retina, Detwiler et al. (1978, 1980) first showed that the time-to-peak responses following the onset of a light flash was shorter in rods further away from a bar of light, indicating that the rod network behaves like a high-pass filter. By introducing an ‘inductive’ pathway in their model, several studies suggested that the activation of a voltage- and time-dependent inward rectifying current (Ih) (Attwell & Wilson, 1980; Hestrin, 1987) might be responsible for the rod high-pass filtering property (Detwiler et al. 1978, 1980; Torre & Owen, 1983; Attwell et al. 1984). Other studies, however, suggest that rod responses are band-pass filtered during light transmission (Attwell, 1986; Armstrong-Gold & Rieke, 2003), presumably due to the membrane capacitive property (Attwell, 1986). Nevertheless, it is uncertain whether the gap junctions between two adjacent rods would contribute to the filtering properties of the rod network, and therefore attenuate the kinetics of rod signals. Recent studies in other retinal neurones and in the neurones in other parts of the central nervous system (CNS) indicate that such a coupling pathway may behave as a low pass filter (Veruki & Hartveit, 2002a,b; Nolan et al. 1999; Galarreta & Hestrin, 2001).
, 百拇医药
Our previous study investigated the cellular localization of gap junction proteins in the salamander retina with a fish connexin35 (Cx35) antibody (that also recognizes murine connexin36, and thus named Cx35/36 in this study). Cx35 was first cloned from the fish retina (O'Brien et al. 1996), and later a homologous murine Cx36 was identified (Condorelli et al. 1998; Shl et al. 1998) and localized in the mammalian retina (Feigenspan et al. 2001; Mills et al. 2001; Deans et al. 2002). We found that Cx35/36 was localized to rod photoreceptors in the salamander retina (Zhang & Wu, 2004), where the ultrastructure of gap junctions has been identified (Custer, 1973; Mariani, 1986). Nonetheless, whether the biophysical profiles of rod electrical coupling agree with the physiological properties of Cx35/36 examined in in vitro systems (White et al. 1999; Srinivas et al. 1999; Teubner et al. 2000) is unknown. Thus it is of great interest to compare the physiology of salamander rod electrical coupling with the Cx35/36 gap junctional properties identified in in vitro (Srinivas et al. 1999; White et al. 1999; Teubner et al. 2000).
, http://www.100md.com
In order to elucidate the biophysical properties of rod electrical coupling, we used the dual whole-cell voltage- and current-clamp recording techniques to directly measure the junctional conductance and the coupling coefficient between paired rods in tiger salamander retinal slices, and also to determine at quantitative levels the strength and dynamics of rod–rod coupling. Here we show that the conductivity of amphibian rod gap junctions is smaller than previously estimated with an average conductance of about 500 pS and an average coupling coefficient (K) of 0.07. Our experimental results, combined with the modelling of the electrically coupled rod network, also suggest that rods behave like a band-pass filter and that the gap junctions between rods do not contribute to the filtering mechanisms of the rod network. In addition, we also compared several attributes of rod electrical coupling with the physiological properties of gene-encoded Cx35/36 gap junctions examined in other in vitro studies.
, http://www.100md.com
Methods
Retinal slice preparation
Larval tiger salamanders (Ambystoma tigrinum) purchased from Charles D. Sullivan, Co. (Nashville, TN, USA) and KON's Scientific Co. Inc. (Germantown, WI, USA) were used in this study. The University Committee on Animal Use at Baylor College of Medicine approved the use of animals and all animals were treated in accordance with the NIH guidelines. The dissection procedures have been previously described (Werblin, 1978; Wu, 1987). In brief, the animals were maintained on a daily 12-h light–dark cycle. The animals were decapitated and the eyes were enucleated and hemisected. The cornea, lens and vitreous were carefully removed. The retina was then removed from the posterior eyecups and was flattened, photoreceptor side up, on filter paper. The retina and filter paper were sectioned into 200–300 μm thick slices. Slices were maintained in Ringer solution at room temperature (22°C).
, http://www.100md.com
Dual whole-cell recordings
The extracellular Ringer solution consisted of (mM): 111 NaCl, 2.5 KCl, 1.8 CaCl2, 1 MgCl2, 10 dextrose, and 5 Hepes, and the pH was adjusted to 7.8 with NaOH. In some experiments, 40 mM tetraethyl ammonium chloride (TEA-Cl) replaced equimolar NaCl in the extracellular medium to block voltage-dependent outward-rectified Ik. There was no significant difference in the averaged gap junctional conductance between the two groups (with or without TEA). The electrodes were pulled from borosilicate glass (TW150F-4, World Precision Instruments, Sarasota, FL, USA) using a Flaming/Brown P-87 puller (Sutter Instrument, Novato, CA, USA), and they had a resistance of 5–7 M when filled with an intracellular solution containing the following (mM): 106 potassium gluconate, 5 NaCl, 2 MgCl2, 5 EGTA and 5 Hepes, adjusted to pH 7.4 with KOH.
, http://www.100md.com
The experiments were performed in dim room lighting. Dual whole-cell recordings were obtained from paired salamander rods with an EPC9/2 amplifier (HEKA Elektronik, Lambrecht, Germany) in voltage- or current-clamp mode. Most of the pipette capacitance was neutralized in the cell-attached configuration using the ‘Cfast’ capacitance neutralization network built into the EPC-9. After the whole-cell configuration was established, the membrane potential of both rods was initially held at –40 mV. Electrical coupling between rods was measured by applying a series of voltage or current step commands to one rod (the driver cell) in order to elicit a membrane current or voltage response in the adjacent rod (the follower cell). Current and voltage signals were filtered at 3 kHz and digitized at 2–5 kHz using a Pentium computer equipped with an ITC-16 data acquisition board (HEKA Elecktronik). Pulse software (v8.65, HEKA Elecktronik) was used to generate voltage and current commands.
, 百拇医药
Data analysis was carried out using PulseFit (v8.65, HEKA Elecktronik), Igor Pro (v4.08, WaveMetrics, Lake Oswego, OR, USA), Excel 2000 (Microsoft), and Origin (v7.0, OriginLab Corp., Northampton, MA, USA). The junctional conductance (Gj) and the coupling coefficient (K) were estimated by calculating the slope of the linear regression curve used to fit the original data. The equations were defined as follows:
and
where A was the offset of current or voltage, Ij was the transjunctional current measured from the follower cell, and Vj was the transjunctional voltage (also defined as Vj=V=Vf–Vd, Vf being the membrane voltage of the follower cell, and Vd being the membrane voltage of the driver cell). All statistic values were given as the mean ± standard deviation and n is the number of cells unless elsewhere indicated.
, http://www.100md.com
Immunocytochemistry
The salamander retinas were fixed in fresh 4% paraformaldehyde/phosphate buffered saline (PBS), pH 7.8, for 30–60 min at room temperature. Following fixation, they were rinsed extensively with PBS. For double-label experiments, the pieces of free-floating vibratome sections were blocked with 3% donkey serum in PBS with 0.5% Triton X-100–0.1% sodium azide for 2 h to overnight in order to reduce non-specific labelling. The tissues were then incubated in a mixture of primary antibodies in the presence of 1% donkey serum–PBS–0.5% Triton X-100–0.1% sodium azide for 3–5 days at 4°C. Controls lacking primary antibodies were blank. Following extensive washes with PBS containing 0.5% Triton X-100–0.1% sodium azide, immunoreactivity was revealed by overnight incubation with immunofluorescent secondary antibodies conjugated to appropriated fluorochromes. After extensive rinsing, the tissues were mounted with Vectashield (Vector Laboratories, Inc., Burlingame, CA, USA) and observed with a confocal laser-scanning microscope (Zeiss LSM 510, Carl Zeiss, Inc., Thornwood, NY, USA). Images were acquired using a 40x or 63x oil-immersion objective lens and Zeiss LSM-PC software. Intensity and size of the images were adjusted using Adobe Photoshop (v 5.0).
, http://www.100md.com
A mouse monoclonal antibody against Cx35/36 (clone 8F6.2) was obtained from Chemicon International (Temecula, CA, USA). It was used at a dilution of 1: 1000 for immunocytochemistry. A rabbit polyclonal antibody against bovine recoverin (1: 1000) was kindly provided by Dr A. M. Dizhoor (Pennsylvania College Optometry, Elkins Park, PA, USA). Secondary antibodies used in the experiments were donkey antimouse IgG and donkey antirabbit IgG conjugated to CY3 (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) or Alexa 488 (Molecular Probes, Eugene, OR, USA) and used in the dilution of 1: 100 in PBS containing 0.5% Triton X-100–0.1% sodium azide.
, 百拇医药
Results
Rod photoreceptors and their expression of Cx35/36 gap junction proteins
Our previous study has demonstrated the occurrence of Cx35/36 gap junctions in rod photoreceptors of the tiger salamander retina (Zhang & Wu, 2004). In the rod sublayer (red, labelled by recoverin antibodies) of the outer nuclear layer (ONL), Cx35/36-positive plaques (green, arrows) are present between rod cell bodies (Fig. 1A), and are restricted at the level distal to the external limiting membrane (Mariani, 1986). The plaques are colocalized with the rod fins where the ultrastructure of gap junctions was observed (Custer, 1973; Mariani, 1986; Zhang & Wu, 2004). Therefore, in order to characterize the biophysical properties of rod gap junctions, all recordings were restricted to the rod sublayer. In some cases, Lucifer yellow dye was loaded into the cells through two patch pipettes to further identify the morphology and position of the rods. One such example is shown in Fig. 1B.
, 百拇医药
A, Cx35/36 (green) immunoreactivity in the salamander outer retina. Vertical single optical section focusing at the outer nuclear layer (ONL). The retina was double labelled with recoverin antibodies (red). Rods (R, faintly stained) were located in the upper tier, whereas cones (C, strongly stained) were located in the lower tier. The characteristic round and elongated Cx35/36-positive plaques were found between rod somas in the distal portion of the ONL (arrows). Scale bar = 10 μm. B, visualization of a pair of rods simultaneously patch clamped and filled with Lucifer yellow through two recoding pipettes.
, 百拇医药
In an earlier electron microscopic study, Lasansky (1973) demonstrated that the salamander rods do not make ribbon junctional contacts to one another, suggesting that there is no excitatory chemical synaptic transmission between rods. It has also been reported that rods do not receive inhibitory inputs from horizontal cells (Lasansky, 1973; Copenhagen & Owen, 1976; Wu, 1992). Thus, our data, along with the others', suggest that rod photoreceptors are mostly interconnected by electrical synapses as opposed to chemical synapses. Therefore, most of our physiological recordings were made in the normal Ringer solution without the addition of blockers to suppress the chemical synaptic transmission.
, 百拇医药
Rod–rod junctional conductance
Simultaneous patch clamp recordings were obtained from a total of 38 pairs of rods. The mean input resistance, measured in the linear region of the I–V relationship (between –40 to –35 mV), was 275 ± 22 M (n= 71). To examine the direct current flow through the electrical synapses of paired rods, we first set the whole-cell configuration in the voltage clamp mode. The membrane potential of two adjacent rods was initially held at –40 mV, near their dark membrane potential (Attwell & Wilson, 1980). A voltage step series (V1) (from –120 to 40 mV with an increment of 20 mV) applied to one rod (driver cell) evoked current responses (I2) in the neighbouring rod (follower cell) of opposite polarity to the responses in the driver cell (I1) (Fig. 2A). This observation confirmed that the rod–rod coupling was preserved in the slice preparations. The appearance of opposite polarity of the transjunctional currents illustrated the presence of a sign-conserving electrical synapse between rod photoreceptors. The relation of transjunctional current (Ij) (measured in the follower cell) and transjunctional voltage (Vj, see Methods) at the steady state (, Fig. 2A) was approximately linear at the membrane potentials tested (Fig. 2B). Switching the driver/follower cell positions resulted in similar current responses and linear relations of Ij–Vj. This property was observed in all pairs of next-neighbouring rods.
, 百拇医药
A, the two paired rods were voltage clamped at –40 mV. When a series of voltage step commands (V1) were applied to cell1 (driver cell), the voltage-activated current responses (I1) were recorded in cell 1 (left panel), and the junctional currents of the opposite polarity (I2) were recorded in cell 2 (follower cell, right panel). B, the plot of transjunctional current (Ij) as a function of transjunctional voltage (Vj) at the steady state () seen in A. C, distribution of junctional conductance for 28 rod–rod pairs. D, the plot of the junctional conductance in each direction. Gj1,2 represents the junctional conductance measured from cell 1 coupled to cell 2, whereas Gj2,1 represents the junctional conductance measured from cell 2 coupled to cell 1. The diagonal line indicates the expected value, showing that the junctional conductance in either direction is similar.
, http://www.100md.com
To further determine the conductivity of rod coupling, we calculated the junctional conductance (Gj). This was determined by the slope of a linear regression-fitting curve (see eqn (1) in Methods) as shown in Fig. 2B. The junctional conductance of the cell in Fig. 2A was 500 pS. A histogram of junctional conductances is given in Fig. 2C. The mean junctional conductance was 500 ± 52 pS (n= 50). The comparison of coupling conductance measured in each direction (i.e. G1,2/G2,1) showed that the conductance was similar and that the electrical synaptic transmission between adjacent rods was reciprocal and symmetrical (Fig. 2D).
, http://www.100md.com
The coupling coefficient of rod–rod electrical synapses
The coupling coefficient (K) of rod–rod electrical synapses can be measured by determining the ratio of the voltage responses of the neighbouring follower cell to that of the driver cell, as shown in the rat retinal studies by Veruki & Hartveit (2002a,b). In the current clamp configuration, injecting a series of negative currents (I1, from –1000 to –100 pA with the increment of 150 pA) into a driver cell evoked a fast transient hyperpolarization followed by a characteristic peak-to-plateau depolarization (V1) in the driver cell (Fig. 3A). The voltage responses in the follower cell (V2) exhibited the same polarity as those measured in V1 (Fig. 3A). Injecting a series of positive currents (from 50 to 650 pA with the increment of 150 pA) into the same driver cell revealed a depolarization in the driver cell and a smaller depolarization in the follower cell (Fig. 3A). Note that the voltage responses of the follower cell exhibited a transient hyperpolarization, which was similar to those of the driver cell. However, the temporal property of the follower cell's transient voltage responses was slower and broader than that of the driver cell (Fig. 3B). The time-to-peak responses of the follower cell (varied with the currents injected into the driver cells) were about 20–100 ms slower than those of the driver cell. Switching the driver/follower cell positions did not change the voltage responses of the driver/follower cells. The activity observed in all pairs of next-neighbouring rods was similar. Measuring the amplitude of the voltage responses at the onset (filled symbols) or the offset (open symbols) of the current injection steps also revealed a rectified relationship between the currents injected to the driver cell and the voltage responses of the driver and follower cells at more depolarizing potentials (Fig. 3C).
, 百拇医药
A, when a series of current step commands (I1) were applied to cell 1, the voltage responses (V1) were recorded in cell 1 (left panel) and the evoked voltage responses (V2) of the same polarity were recorded in cell 2 (right panel). B, the plots of voltage responses of cell 1 and cell 2 corresponding to V1 and V2, respectively, in an expanded time scale. C, the plots of voltage responses of V1 and V2 as a function of currents injected into the driver cell (I1) at the instantaneous state (filled symbols) and the steady state (open symbols). D, the plot of the coupling coefficient in each direction of paired rods. The diagonal line indicates the expected value, showing that the coupling coefficient in either direction is similar. E, the plot of the K ratio versus the ratio of the rod input resistance (see text). The diagonal line indicates the expected value, showing that the K ratio is similar to the ratio of the input resistance.
, http://www.100md.com
The coupling coefficient (K) of rod–rod electrical transmission (see eqn (2) in Methods) measured at the instantaneous state (filled symbols, Fig. 3A) ranged from 0.02 to 0.3 with the mean value of 0.07 ± 0.01 (n= 40). Most cell pairs had symmetrical K-values for both directions (i.e. K1,2 and K2,1) (Fig. 3D). In some pairs, the coupling appeared to be rectified in spite of the fact that the junctional conductance (Gj) for each direction was symmetrical. When the mean K ratio (the larger K-value in either direction divided by the smaller K-value in the opposite direction) was calculated (2.20 ± 0.49, n= 20 cell pairs), it was found to be very close to the mean ratio of input resistances of paired rods (2.15 ± 0.45, n= 20 cell pairs) (the larger input resistance divided by the smaller input resistance). Plotting the K ratio as a function of the ratio of input resistances revealed the correlated relationship (Fig. 3E). This suggested that the rectification occurring in rod–rod coupling was due to a disparity between the membrane input resistances of the coupled rods. These results are similar to a previous study of the rat retina (Veruki & Hartveit, 2002a,b).
, 百拇医药
The voltage independency of rod–rod electrical transmission
The measurement of the gap junctional conductance demonstrated that the rod–rod coupling conductivity exhibited weak voltage sensitivity within the physiological range of rod photoreceptors (see Fig. 2B). This implies that the reciprocal communication between rods is independent of the transjunctional voltage, especially when rods are illuminated uniformly in the dark when their membrane potential is usually at –40 mV. But it is not known whether the efficient transmission between rods is affected when the resting membrane potentials of two rods differ. To test this idea, two adjacent rods were current clamped at 0 pA. As shown in Fig. 4A, a sinusoidal current pulse of 5 Hz (I1) applied to the driver cell evoked a peak-to-peak voltage response (V2) of 60 mV in the follower cell at its resting membrane potential. When the follower cell resting membrane potential was moderately hyperpolarized or depolarized by injecting a small negative or positive current, the amplitude of the follower cell voltage responses elicited by the current injection to the driver cell was less affected. However, when the resting membrane potential of the follower cell was either hyperpolarized to about –70 mV or depolarized to about –10 mV by injecting a current of ± 280 pA (I2), and the same sinusoidal current pulse of 5 Hz was applied to the driver cell, the amplitude of the follower cell voltage responses was partially suppressed. As the follower cell was hyperpolarized to more than –100 mV or depolarized to more than 0 mV, the same sinusoidal current pulse applied to the driver cell elicited much less voltage responses in the follower cell. When the normalized coupling coefficient (K) (to that determined at 0 pA holding current) was plotted as a function of the currents injected into the follower cell (Fig. 5B) and was fitted by the Gaussian curve, it showed a bell-shaped relation, reflecting that within a certain range of current injections (between –300 and 50 pA), the normalized K remained relatively constant. Outside this range, the normalized K decreased. Yet, a large amount of the residual gap junctional current still remained. The same observation was found in three out of three rod pairs.
, http://www.100md.com
A, the effect of varying membrane potentials on the rod electrical transmission. When the membrane potential of cell 2 was set at different voltages by injecting varying currents (I2), a sinusoidal current of 5 Hz (I1) applied to cell 1 elicited voltage responses in cell2 (V2). B, the plot of normalized coupling coefficient (K) (to that determined at 0 pA holding current) at each given I2. The curve was fitted with the Gaussian equation. C, the effect of varying membrane potentials on rod junctional conductance. A series of voltage step commands applied to cell 1 (V1) activated junctional currents in cell 2 (I2) at different holding potentials (V2 h: 0 mV, –40 mV and –80 mV). D, the plot of the Ij–Vj relation at each given holding potential. The conductance was determined by the slope of the Ij–Vj curve. E, the plot of the normalized conductance (to that determined at –40 mV holding potential) as the function of the holding potential of cell 2. The curve was fitted with the Boltzmann equation. The holding potential that produced half conductance was estimated at –2.66 mV.
, http://www.100md.com
A, the sinusoidal current stimuli (I1) of varying frequencies applied to cell 1 elicited voltage responses in cell 1 (V1, grey curve) and in cell 2 (V2, black curve). B, the plots of the normalized V1 (), V2 (), K () (left axis), and the degree of phase shift (, right axis) as the function of current frequency. The cut-off frequency was estimated as 33 Hz. C, the sinusoidal voltage stimuli (V1) of varying frequencies applied to cell1 activated transjunctional currents in cell 2 (I2). D, the plot of the normalized transjunctional current (I/Imax) as the function of voltage frequency.
, http://www.100md.com
Alternatively, to probe how the rod resting membrane potentials affect the rod coupling conductance, two adjacent rods were simultaneously recorded in the voltage-clamp configuration (Fig. 4C). The complete traces of current responses of the follower cell (I2) (while its membrane potential was held at 0, –40 and –80 mV) are depicted in Fig. 4C. At any given holding potential between –80 and 40 mV, the Ij–Vj relation was approximately linear (Fig. 4D). However, it should be pointed out that when the membrane potential of the follower cell was held at a potential more negative than –80 mV or more positive than 40 mV, it was difficult to measure the junctional current in the follower cells. Thus, we only determined the junctional conductance (Gj) from the slope of the Ij–Vj curve within the range of –80 to 40 mV. Then we normalized it to the conductance obtained at the –40 mV, and fitted the data points with the Boltzmann equation. We found that, at more positive holding potentials than –20 mV, the gap junctional conductance started to decrease (Fig. 4E). This finding is consistent with the results obtained in the current clamp mode. Nonetheless, while the holding potential was below –30 mV, the junctional conductance was almost independent of the rod membrane potential. The voltage that produced the half-maximal conductance was estimated to be –2.66 mV (n= 4 cell pairs). These results are consistent with above current-clamp data, suggesting that within its physiological range, the rod membrane potential would not attenuate rod electrical signals significantly.
, 百拇医药
The characteristics of frequency dependence of the rod network and gap junctions
It has been suggested that the coupled rod network acts as a band-pass filter in shaping the visual signal transmitting to the second-order cells (Attwell, 1986; Armstrong-Gold & Rieke, 2003). Yet it is not clear if the coupling pathway (or gap junctions) between rods contributes to the filtering properties of the rod network. We therefore recorded rod pairs in a whole-cell configuration that was set in the current clamp mode as shown in Fig. 5A. The sinusoidal current stimuli (I1) of varying frequencies (1–100 Hz) applied to the driver cell evoked the voltage responses of the driver cell (V1, grey curve) and the follower cell (V2, black curve). As illustrated in Fig. 5B, averaged from five rod pairs, we found that the normalized voltage response of the driver cells (V1, open triangles) was 94%, 100%, and 96% at 1 Hz, 2 Hz, and 5 Hz, respectively, whereas the normalized voltage response of the follower cell (V2, filled triangles) was 91% at 1 Hz, 100% at 2 Hz, and 87% at 5 Hz. When the frequency of the sinusoidal current stimuli (I1) increased continuously, the normalized voltage responses of both the driver and the follower cells decreased, with the latter one decreasing significantly. Similarly, the normalized coupling coefficient KN (filled circles) was 94% at 1 Hz, 100% at 2 Hz and 88% at 5 Hz, and it gradually reduced in amplitude and developed a phase-shift (open circles) over the frequency range of 5 to 100 Hz (Fig. 5B). Thus, at the low frequency (such as at 1 Hz), voltage signals of both the driver and follower cell were attenuated, and then they peaked at a frequency of about 2 Hz. In contrast, the high frequency voltage signals passing through gap junctions to the follower cell were also largely attenuated with the cut-off frequency (defined as the frequency that produced the half-maximal KN value) measured at 33 Hz. Since the attenuation of voltage responses over a wide range of temporal frequency was found not only in the follower cell but also in the driver cell, it is implied that rod voltage responses are shaped before they are transmitted to adjacent rods.
, 百拇医药
To further determine the frequency dependence of rod gap junctions, we voltage clamped the membrane potential of the follower cell (V2) at –40 mV and manipulated the membrane potential of the driver cell (V1) by applying sinusoidal voltage pulses of varying frequency. As illustrated in Fig. 5C, when the frequency of sinusoidal voltage stimuli (V1) was increased, the amplitude of the current (I2) passing through gap junctions to the follower cell did not significantly decrease over the frequency range of 1–50 Hz. Averaged from five rod pairs, we observed that, at the frequency of 50 Hz, the normalized current response was reduced by only 10 ± 3% of the maximum current measured at the frequency of 1 Hz (Fig. 5D). In contrast, the normalized KN at the frequency of 50 Hz was reduced by 57 ± 10% of the maximum KN determined at the frequency of 2 Hz. Therefore, the measurement of current responses as a function of the frequency was in direct contradiction to the measurement of voltage responses as a function of the frequency. This suggests that the higher frequency filtering property observed here most likely reflects the RC property of the rod network.
, 百拇医药
In order to confirm whether rod gap junctions behave as a resistor, and to examine how RC properties of the rod membrane are related to the filtering properties, we analysed a simplified model of the rod network (Fig. 6A). In this model, the membrane resistance (R1 (or 1/G1) and R2 (or 1/G2)) and capacitances (C) represent the lumped equivalent of the resistance and capacitance across driver (1) and follower (2) rod membranes, while two rods are connected through a resistive pathway Rj (= 1/Gj). Assuming Ih is ignored, and therefore the high pass filtering pathway is not considered here, the equation defining the relationship of coupling coefficient (K) as the function of frequency can be expressed as
, http://www.100md.com
At a given capacitance of 40 pF (Attwell et al. 1984) and an input resistance of 250 M (obtained from this study), the amplitude of rod signals spreading to the adjacent rods is frequency dependent (Fig. 6B). The cut-off frequency calculated from this model is 32 Hz (*, Fig. 6C), which is very close to what we measured experimentally. In addition, considering the rod input resistance variation from 100 to 500 M, the normalized voltage responses at a given higher frequency (to the steady state responses) would be attenuated significantly. The gap junctional conductance affecting the frequency-dependent voltage attenuation would be much less spectacular than that of rod membrane input resistance (Fig. 6B and C). Therefore, the agreement of the physiological data and the mathematical calculations supports the notion that rod gap junctions behave in a linear (ohmic) manner in mediating rod–rod electrical coupling. The frequency-dependent attenuation of rod voltage responses is not surprising given the rod plasma membrane capacitive and resistive properties.
, 百拇医药
A, a schematic diagram of the equivalent circuit of the rod pair. The two next neighbouring rods are interconnected by a resistance of Rj (= 1/Gj). R1, R2 and C represent the lumped equivalent of the resistance and capacitance across each rod membrane in the pairs, where R1= 1/G1 and R2= 1/G2 (G1 and G2 are the conductance of cell 1 and cell 2, respectively). B, the computer simulated normalized coupling coefficient (KN) changes as a function of frequency (when Gj and R2 were manipulated using different values). C, the plot of cut-off frequency (see eqn (3) in Results) as the function of junctional conductance at each given membrane input resistance. The calculated cut-off frequency is 32 Hz (*) when Gj, R2, and C are 500 pS, 250 M, and 40 pF, respectively.
, http://www.100md.com
The intensity dependence of the coupled rod network
In the salamander retina, rods are able to respond to light ranging in intensity from a few photons to several thousands of photons with characteristic temporal kinetics (Bader et al. 1978; Attwell et al. 1984; Yang & Wu, 1997). While rods respond to dim light with a slow graded hyperpolarization, their responses to brighter light become faster with a typical transient hyperpolarization. It is likely that when rod signals propagate through the rod network, the amplitude of rod voltage responses would not only be limited by the coupling strength of the electrical synapses, but would also be attenuated by the filtering mechanism as the light intensity increases. In order to determine the degree of filtering-induced attenuation at different log units of intensity, we measured the time-to-peak of rod transient responses obtained from intracellular recordings (n= 3) (see Fig. 3, Yang & Wu, 1997; and X. L. Yang & S. M. Wu, unpublished results), and calculated the amplitude of transient voltage responses of the follower cell with/without considering RC properties of the rod membrane. We found that during the brightest light flash, the rise time of this transient response could be less than 50 ms. Our calculations showed that over eight log units of the light intensity span, the dim light responses (log I < (4)) of rods with slow kinetics would efficiently pass through rod gap junctions (, Fig. 7). However, as light intensity increases, the transient rod responses would be brief enough to be filtered (Fig. 5). Therefore, the peak amplitude of rod responses (log I > (4)) would likely be attenuated (, frequency, Fig. 7). Thus the light responses of rods laterally spreading within the network would be intensity dependent.
, http://www.100md.com
The plots of the voltage-intensity relation of the driver cell (), the follower cell without considering filtering properties of the rod network (), and the follower cell having peak responses filtered by the rod network () as light intensity increases.
Discussion
The anticipation of Cx35/36 gap junctions in rod electrical coupling
Since we have previously demonstrated that Cx35/36 gap junction proteins are present in an organized pattern in the rod network of the salamander retina (Zhang & Wu, 2004), one aim of this continuing study is to understand whether the physiological property of rod electrical coupling is consistent with the profiles of Cx35/36 gap junctions. Although, to date, the lack of potent and specific Cx35/36 gap junction blockers makes it difficult to pharmacologically distinguish Cx35/36 from other connexin proteins in salamander rods, there are several lines of evidence suggesting that Cx35/36 may play an important role in the rod electrical coupling of this species. First, morphologically, we have shown that among all of the connexin proteins examined, only Cx35/36 is exclusively localized in rod photoreceptors at locations where ultrastructure of gap junctions has been identified (Custer, 1973; Mariani, 1986; Zhang & Wu, 2004). Secondly, the staining pattern of Cx35/36 in the rod network is similar to that of Cx35/36 gene-encoded channels expressed in transfected human HeLa cells (Teubner et al. 2000).
, http://www.100md.com
Thirdly, in the present electrophysiological study, we have demonstrated that the biophysical properties of rod gap junctions resemble the physiology of gene-encoded Cx35/36 gap junctions. Several earlier reports have shown that Cx35/36 gap junctions studied in in vitro expression systems exhibit three characteristics: (1) small channel conductance; (2) voltage independence; and (3) formation of homologous channels (Srinivas et al. 1999; White et al. 1999; Teubner et al. 2000). Similarly, in the salamander retina, the rod coupling exhibited the smaller conductance of 500 pS. Within the rod physiological range, the conductivity of rod gap junctions was independent of the transjunctional voltage between rods, and was also independent of the rod plasma membrane potential. Furthermore, since current flow passed equally well in both directions, this suggests that rod electrical synapses were symmetrical and bi-directional. These are results consistent with the notion that Cx35/36 forms homologous gap junctions between neighbouring neurones (White et al. 1999; Teubner et al. 2000). Interestingly, all of these characteristics have also been observed in Cx35/36-containing neurones studied in in situ brain slices (Landisman et al. 2002; Galarreta & Hestrin, 1999) and in Cx35/36-positive AII amacrine/ON cone bipolar cells studied in rat retinal slices (Veruki & Hartveit, 2002a,b). Therefore, our data indicate that Cx35/36 gap junctions are likely to mediate rod–rod coupling in the salamander retina. However, as selective gap junctional blockers are discovered, further pharmacological and physiological studies are needed to confirm this claim.
, http://www.100md.com
Small gap junctional conductivity mediating salamander rod electrical coupling
Our characterization of rod gap junctional properties revealed a weak coupling of rod photoreceptors. The experimentally measured junctional conductance of 500 pS was much smaller than the previously computer-simulated value of 3.3 nS (Attwell et al. 1984). Yet, it is much closer to the cone–cone coupling conductance found in the mammalian retina (Hornstein et al. 2004; Li & DeVries, 2004). It may be argued that in the slice preparations, the rod network was partially cut off, especially for superficial cell pairs, and that the junctional conductance in vivo might be larger than what we observed here. However, Attwell & Wilson (1980) found that rods were coupled more strongly in the retinal slices than in the flat-mount retina. They speculated that this might be due to the decrease of the number of paths available for current flow away from the site of current injection. Thus the slice preparation itself should not be a significant factor affecting the small rod junctional conductance.
, http://www.100md.com
We have quantitatively studied gap junction patterns between rods. We previously calculated the average number of Cx35/36 positive plaques between paired rods to be 4.63 (Zhang & Wu, 2004). Assuming the unitary channel conductance of Cx35/36 is 10–14 pS (Srinivas et al. 1999) and the coupling conductance between paired rods is 500 pS, we estimated that there should be as many as 35–50 channels (500 x 10–12)/(10–14 x 10–12)) that are open under dark-adapted conditions. Each Cx35/36-positive plaque may therefore accommodate 8–11 activated gap junctions. In a freeze fracture study of gap junctions in the toad retina, Gold & Dowling (1979) showed that the average area per junction in rod–rod pairs was 0.15 ± 0.05 μm2 and the density of junctional particles was 5 x 103μm–2. This suggests that the number of junctional particles in a plaque would be 750 (0.15 x 5 x 103). This also suggests that the activation of 8–11 gap junctions in the dark only represents 1–1.5% (8/750 or 11/750) of the total number of channels in a cluster. This percentage is much smaller than the estimate of 10% of the total number of channels in a cluster assumed to be activated as suggested in a Cx43 study by Bukauskas et al. (2000).
, 百拇医药
The low coupling conductance (500 pS) resulted in a small coupling coefficient of 0.07, which indicates a weak coupling strength between rods. This weak interconnection between rods may be desirable, because the greatest sensitivity of rods is at the effective operational range, i.e. near the dark membrane potential (Attwell & Wilson, 1980; Capovilla et al. 1987; Yang & Wu, 1996, 1997), and the weak coupling ensures a given signal divided into smaller signals that will fall within the sensitive and linear operational range of rods. These small signals are then expected to be transmitted/amplified precisely and maximally to the second-order cells by means of elaborate chemical synapses. Likewise, in the CNS (including neocortex, cerebellum and thalamus), this similar coupling coefficient has been shown to depolarize neighbouring interneurones to a subthreshold potential, thereby facilitating the synchronized firing of action potentials (reviewed in Galarreta & Hestrin, 2001). Therefore, we think that the small conductivity of rod gap junctions may be necessary to maintain rods in the optimal state for integrating incoming small signals within a narrow operational range.
, 百拇医药
The significance of the linearity of rod gap junctions
The voltage independence of rod gap junctions within the rod physiological range combined with the frequency independence of rod junctional currents strongly implies that the rod–rod coupling pathway behaves as a linear resistor. This feature may provide a mechanism to prevent rod uncoupling within the rods' dynamic range and also to improve the signal-to-noise ratio within the whole intensity range of light illumination. While voltage would not be the primary modulator of rod gap junctions, it is possible that they would be modulated by neuromodulators and/or intracellular second messengers. In other electrically coupled retinal neurones, such as horizontal and amacrine cells, dopamine, cAMP, and nitric oxide have been found to modulate the gap junctional conductivity of these cells (Lasater, 1987; Hampson et al. 1992; Mills & Massey, 1995; Xin & Bloomfield, 2000). Correspondingly, this may be true in rod electrical coupling. A recent report showed that Cx35 hemi-channels expressed in Xenopus oocytes were modulated by the cAMP-mediated protein kinase A (PKA) pathway (Mitropoulou & Bruzzone, 2003). The consensus sequence sites for PKA phosphorylation have been identified in Cx35 proteins (O'Brien et al. 1996). Furthermore, the dye diffusion coefficient in cells expressing Cx35 was modulated by PKA (O'Brien et al. 2004), and Cx36 associated proteins were found to be phosphorylated by cAMP (Sitaramayya et al. 2003). Therefore, if Cx35/36 is functionally involved in rod electrical coupling, rod electrical synapses are anticipated to be modulated at the cellular and molecular levels. These synapses may also be targeted by second messenger mediated-signalling transduction pathways or by light-dependent signalling pathways. Further pharmacological and physiological studies of the modulation of rod–rod coupling are needed in order to confirm this hypothesis.
, http://www.100md.com
The temporal filtering property of the rod network shapes intensity-dependent rod signals
In this study, as shown in Fig. 6, we demonstrated that the low-pass filtering of higher frequency rod signals might be attributed to the RC properties of the rod plasma membrane. This conclusion added a piece of additional evidence to the mechanisms of the band-pass filtering property of the rod network (Detwiler et al. 1978, 1980; Torre & Owen, 1983; Attwell et al. 1984). We found that voltage responses of coupled rods were gradually attenuated as the frequency of current stimuli increased beyond 5 Hz. This indicates that rod electrical synapses favour the transmission of slow potential changes and it agrees with the previous findings of frequency dependence of Cx35/36-containing retinal neurones and brain interneurones (Veruki & Hartveit, 2002a,b; Galarreta & Hestrin, 1999; Nolan et al. 1999; Landisman et al. 2002). One possible explanation for this finding is that Cx35/36 gap junctions might impart a ‘low-pass’ filtering property (Veruki & Hartveit, 2002a,b; Nolan et al. 1999; Galarreta & Hestrin, 2001). However, we found that the transjunctional current under voltage clamp conditions was not significantly altered as the frequency increased (Fig. 5C), indicating that DC coupling is attributed to rod gap junctions. Thus, our data argue that the attenuation of voltage responses of coupled rods by signal frequencies is not the result of the frequency-dependent processes in the gap junctions between rods. Instead it reflects rod membrane RC filtering properties. This argument is supported by our mathematical calculations in which we showed that the cut-off frequency (32 Hz) was very close to the value (33 Hz) measured experimentally, which is a result consistent with the notion that gap junctions in salamander rods behave in a linear manner. A similar observation was also made in the study of primate cone–cone coupling (Hornstein et al. 2004).
, http://www.100md.com
Rod photoreceptors are non-spiking neurones in that they respond to light with graded hyperpolarization (Schwartz, 1976; Yang & Wu, 1997). The current response to a single photon is about 2 pA with a rise time of about 2 s (Baylor et al. 1979; Baylor & Nunn, 1986). When light becomes brighter, rod light responses become larger with a faster rise time (Bader et al. 1978; Yang & Wu, 1997). For a saturating light flash, the peak current response is about 50–100 pA and with a rapid rise time of 20–50 ms (Baylor & Nunn, 1986). The frequency-dependent attenuation of the coupled rod network (Fig. 7) allows small signals with slow kinetics to spread to the neighbouring rod without much attenuation, whereas large responses with faster kinetics could be attenuated. This may be a mechanism used by the retina to suppress rod inputs to second-order cells during bright light so that the cone input can take a more dominant role.
, 百拇医药
References
Armstrong-Gold CE & Rieke F (2003). Bandpass filtering at the rod to second-order cell synapse in salamander (Ambystoma tigrinum) retina. J Neurosci 23, 3796–3806.
Attwell D (1986). The Sharpey-Schafer lecture. Ion channels and signal processing in the outer retina. Q J Exp Physiol 71, 497–536.
Attwell D & Wilson M (1980). Behaviour of the rod network in the tiger salamander retina mediated by membrane properties of individual rods. J Physiol 309, 287–315.
, 百拇医药
Attwell D, Wilson M & Wu SM (1984). A quantitative analysis of interactions between photoreceptors in the salamander (Ambystoma) retina. J Physiol 352, 703–737.
Attwell D, Wilson M & Wu SM (1985). The effect of light on the spread of signals through the rod network of the salamander retina. Brain Res 343, 79–88.
Bader CR, MacLeish PR & Schwartz EA (1978). Responses to light of solitary rod photoreceptors isolated from tiger salamander retina. Proc Natl Acad Sci U S A 75, 3507–3511.
, http://www.100md.com
Baylor DA, Fuortes MG & O'Bryan PM (1971). Receptive fields of cones in the retina of the turtle. J Physiol 214, 265–294.
Baylor DA, Lamb TD & Yau KW (1979). Responses of retinal rods to single photons. J Physiol 288, 613–634.
Baylor DA & Nunn BJ (1986). Electrical properties of the light-sensitive conductance of rods of the salamander Ambystoma tigrinum. J Physiol 371, 115–145.
Belgum JH & Copenhagen DR (1988). Synaptic transfer of rod signals to horizontal and bipolar cells in the retina of the toad (Bufo marinus). J Physiol 396, 225–245.
, http://www.100md.com
Bukauskas FF, Jordan K, Bukauskiene A, Bennett MV, Lampe PD, Laird DW & Verselis VK (2000). Clustering of connexin 43-enhanced green fluorescent protein gap junction channels and functional coupling in living cells. Proc Natl Acad Sci U S A 97, 2556–2561.
Capovilla M, Hare WA & Owen WG (1987). Voltage gain of signal transfer from retinal rods to bipolar cells in the tiger salamander. J Physiol 391, 125–140.
Condorelli DF, Parenti R, Spinella F, Trovato Salinaro A, Belluardo N, Cardile V & Cicirata F (1998). Cloning of a new gap junction gene (Cx36) highly expressed in mammalian brain neurons. Eur J Neurosci 10, 1202–1208.
, 百拇医药
Copenhagen DR & Owen WG (1976). Coupling between rod photoreceptors in a vertebrate retina. Nature 260, 57–59.
Custer NV (1973). Structurally specialized contacts between the photoreceptors of the retina of the axolotl. J Comp Neurol 151, 35–56.
Deans MR, Volgyi B, Goodenough DA, Bloomfield SA & Paul DL (2002). Connexin36 is essential for transmission of rod-mediated visual signals in the mammalian retina. Neuron 36, 703–712.
, 百拇医药
Detwiler PB, Hodgkin AL & McNaughton PA (1978). A surprising property of electrical spread in the network of rods in the turtle's retina. Nature 274, 562–565.
Detwiler PB, Hodgkin AL & McNaughton PA (1980). Temporal and spatial characteristics of the voltage response of rods in the retina of the snapping turtle. J Physiol 300, 213–250.
Feigenspan A, Teubner B, Willecke K & Weiler R (2001). Expression of neuronal connexin36 in AII amacrine cells of the mammalian retina. J Neurosci 21, 230–239.
, 百拇医药
Galarreta M & Hestrin S (1999). A network of fast-spiking cells in the neocortex connected by electrical synapses. Nature 402, 72–75.
Galarreta M & Hestrin S (2001). Electrical synapses between GABA-releasing interneurons. Nat Rev Neurosc 2, 425–433.
Gold GH & Dowling JE (1979). Photoreceptor coupling in retina of the toad, Bufo marinus. I. Anatomy. J Neurophysiol 42, 292–310.
Hampson EC, Vaney DI & Weiler R (1992). Dopaminergic modulation of gap junction permeability between amacrine cells in mammalian retina. J Neurosci 12, 4911–4922.
, http://www.100md.com
Hestrin S (1987). The properties and function of inward rectification in rod photoreceptors of the tiger salamander. J Physiol 390, 319–333.
Hornstein EP, Verweij J & Schnapf JL (2004). Electrical coupling between red and green cones in primate retina. Nat Neurosci 7, 745–750.
Lamb TD & Simon EJ (1976). The relation between intercellular coupling and electrical noise in turtle photoreceptors. J Physiol 263, 257–286.
, http://www.100md.com
Landisman CE, Long MA, Beierlein M, Deans MR, Paul DL & Connors BW (2002). Electrical synapses in the thalamic reticular nucleus. J Neurosci 22, 1002–1009.
Lasansky A (1973). Organization of the outer synaptic layer in the retina of the larval tiger salamander. Philos Trans R Soc Lond B Biol Sci 265, 471–489.
Lasater EM (1987). Retinal horizontal cell gap junctional conductance is modulated by dopamine through a cyclic AMP-dependent protein kinase. Proc Natl Acad Sci U S A 84, 7319–7323.
, 百拇医药
Li W & DeVries SH (2004). Separate blue and green cone networks in the mammalian retina. Nat Neurosci 7, 75175–75176.
Mariani AP (1986). Photoreceptors of the larval tiger salamander retina. Proc R Soc Lond B Biol Sci 227, 483–492.
Mills SL & Massey SC (1995). Differential properties of two gap junctional pathways made by AII amacrine cells. Nature 377, 734–737.
Mills SL, O'Brien JJ, Li W, O'Brien J & Massey SC (2001). Rod pathways in the mammalian retina use connexin 36. J Comp Neurol 436, 336–350.
, 百拇医药
Mitropoulou G & Bruzzone R (2003). Modulation of perch connexin35 hemi-channels by cyclic AMP requires a protein kinase A phosphorylation site. J Neurosci Res 72, 147–157.
Nolan MF, Logan SD & Spanswick D (1999). Electrophysiological properties of electrical synapses between rat sympathetic preganglionic neurones in vitro. J Physiol 519, 753–764.
O'Brien J, Al-Ubaidi MR & Ripps H (1996). Connexin 35: a gap-junctional protein expressed preferentially in the skate retina. Mol Biol Cell 7, 233–243.
, http://www.100md.com
O'Brien J, Nguyen HB & Mills SL (2004). Cone photoreceptors in bass retina use two connexins to mediate electrical coupling. J Neurosci 24, 5632–5642.
Schwartz EA (1976). Electrical properties of the rod syncytium in the retina of the turtle. J Physiol 257, 379–406.
Sitaramayya A, Crabb JW, Matesic DF, Margulis A, Singh V, Pulukuri S & Dang L (2003). Connexin 36 in bovine retina: lack of phosphorylation but evidence for association with phosphorylated proteins. Vis Neurosci 20, 385–395.
, 百拇医药
Shl G, Degen J, Teubner B & Willecke K (1998). The murine gap junction gene connexin36 is highly expressed in mouse retina and regulated during brain development. FEBS Lett 428, 27–31.
Srinivas M, Rozental R, Kojima T, Dermietzel R, Mehler M, Condorelli DF, Kessler JA & Spray DC (1999). Functional properties of channels formed by the neuronal gap junction protein connexin36. J Neurosci 19, 9848–9855.
Teubner B, Degen J, Sohl G, Guldenagel M, Bukauskas FF, Trexler EB, Verselis VK, De Zeeuw CI, Lee CG, Kozak CA, Petrasch-Parwez E, Dermietzel R & Willecke K (2000). Functional expression of the murine connexin 36 gene coding for a neuron-specific gap junctional protein. J Membr Biol 176, 249–262.
, 百拇医药
Torre V & Owen WG (1983). High-pass filtering of small signals by the rod network in the retina of the toad, Bufo marinus. Biophys J 41, 305–324.
Veruki ML & Hartveit E (2002a). AII (Rod) amacrine cells form a network of electrically coupled interneurons in the mammalian retina. Neuron 33, 935–946.
Veruki ML & Hartveit E (2002b). Electrical synapses mediate signal transmission in the rod pathway of the mammalian retina. J Neurosci 22, 10558–10566.
, 百拇医药
Werblin FS (1978). Transmission along and between rods in the tiger salamander retina. J Physiol 280, 449–470.
White TW, Deans MR, O'Brien J, Al-Ubaidi MR, Goodenough DA, Ripps H & Bruzzone R (1999). Functional characteristics of skate connexin35, a member of the gamma subfamily of connexins expressed in the vertebrate retina. Eur J Neurosci 11, 1883–1890.
Wu SM (1987). Synaptic connections between neurons in living slices of the larval tiger salamander retina. J Neurosci Meth 20, 139–149.
, http://www.100md.com
Wu SM (1992). Feedback connections and operation of the outer plexiform layer of the retina. Curr Opin Neurobiol 2, 462–468.
Wu SM (1994). Synaptic transmission in the outer retina. Annu Rev Physiol 56, 141–168.
Xin D & Bloomfield SA (2000). Effects of nitric oxide on horizontal cells in the rabbit retina. Vis Neurosci 17, 799–811.
Yang XL & Wu SM (1996). Response sensitivity and voltage gain of the rod- and cone-horizontal cell synapses in dark- and light-adapted tiger salamander retina. J Neurophysiol 76, 3863–3874.
, http://www.100md.com
Yang XL & Wu SM (1997). Response sensitivity and voltage gain of the rod- and cone-bipolar cell synapses in dark-adapted tiger salamander retina. J Neurophysiol 78, 2662–2673.
Zhang J & Wu SM (2004). Connexin35/36 gap junction proteins are expressed in photoreceptors of the tiger salamander retina. J Comp Neurol 470, 1–12., 百拇医药(Jian Zhang and Samuel M. )